Recently, the International Data Corporation (IDC) released the first quarter of 2024 China's mobile phone market tracking report, Honor won the first place with a market share of 17.1%, Huawei followed closely behind, OPPO, Apple and vivo ranked third to fifth respectively.

Among them, Honor's forward-looking layout and technology leadership in the AI field have become one of the key factors driving the sales of Honor's high-end flagship products. According to Wang Jiping, thanks to the blessing of AI functions, as well as all-round upgrades such as images and screens, the shipments of the Honor Magic6 series in the first quarter exceeded the sum of the shipments in the first two quarters of the previous generation.

For the market, 2024 seems to be the first year of AI, and mobile phone manufacturers must talk about AI, but for Honor, as the first mobile phone manufacturer in the industry to put forward the concept of mobile phone AI, the first year of AI began in 2016, and platform-level AI is the direction it continues to adhere to.

Mobile phone AI≠ AI mobile phone

What is an AI phone?

In a nutshell: the essence of AI mobile phones is an agent with autonomous perception, autonomous decision-making and execution, and digital personality. AI mobile phones are subversive in definition, function, and user experience.

How do AI smartphones differ from the previous generation of smartphones?

The previous generation of smartphones was already equipped with some AI features, such as the 2019 Apple on the iPhone 11 on the launch of the AI photo processing technology Deep Fusion, which uses machine learning to capture photos in low to medium light, takes nine shots in a row and finally compothesizes high-detail, low-noise images. The intelligent recommendation of shopping platforms and short video platforms also relies on AI algorithms. It can be said that the previous generation of smart phones has the ability to run end-side AI on both the hardware side and the software side.

So, what's new about the new generation of AI smartphones?

By definition, next-gen AI smartphones refer to smartphones with NPU computing power greater than 30 TOPS, SoCs equipped with faster and more efficient Gen AI models on the device side, and Gen AI models including Stable Diffusion and various large language models running on the device side. According to IDC's classification, the previous generation of smartphones equipped with preliminary AI functions can be collectively referred to as "hardware-enabled AI phones", and their NPU TOPS≤30。 Take the iPhone released by Apple in 2017 X, for example, is equipped with the A11 The BIONIC Chip Neural Network Engine can perform up to 600 billion operations per second, which is capable of machine learning tasks, and the face recognition function can be adjusted and adapted to the changes in the appearance of the owner. The new generation of AI smartphones has achieved comprehensive improvements in the speed, efficiency, application scenarios, and chip computing power of the device-side AI.

The new generation of AI mobile phones should have the ability to efficiently utilize computing power, perceive the real world, learn from themselves, and create. According to OPPO and IDC, in terms of efficient utilization of computing power, AI mobile phones need to make efficient use of computing resources to meet the needs of generative AI computing and device-side deployment. In terms of self-learning ability, in addition to the learning ability of the model itself, the AI mobile phone continuously understands the user's habits through machine learning and improves the interactive experience through self-learning ability, and in terms of creative ability, the AI mobile phone needs to be creative, and the generative service provides users with continuous inspiration and knowledge support.

AI mobile phones are expected to truly become a comfortable interaction

Compared with traditional smartphones, AI mobile phones bring users a new human-computer interaction experience through personalization, memory, perception and management capabilities, breaking the boundaries between systems, services, devices and people.

On the one hand, AI mobile phones can provide accurate and customized suggestions and answers based on accurate semantic understanding capabilities and huge knowledge data graphs, superimposed on the learning of users' personalized information and habits.

On the other hand, the ultimate goal of AI mobile phones is to get rid of bloated and numerous apps, vertically integrate end-to-end applications, and deeply connect independent islands.

Possess the ability to make independent decisions

AI phones are not only equipped with generative AI functions, but also play the role of AI The role of the agent. From the perspective of the industry, AI phones do not only refer to mobile phones with generative AI capabilities. It is difficult for independent large models to support other applications on mobile phones, and each app is still an independent island, but integrating these models into the mobile phone system can break down the barriers between applications and improve the system's capabilities to AI At the agent level, not only can the large model be blessed with intelligent assistants, but also the system tools can also have AIGC capabilities. According to Zhao Ming, CEO of Honor, AI mobile phones are a more basic and systematic AI computing platform.

AI mobile phones not only have the needs of the underlying hardware, but also have the innovation of AI and interactive experience at the OS level, which is an AI in itself Agent。 They can perceive the needs of users, make independent decisions and implementation, and even learn user habits to provide users with more personalized services. According to the personal LLM agent review paper jointly released by Xiaomi, Huawei, vivo and other companies and the Intelligent Industry Research Institute of Tsinghua University, AI Agents can be classified into five levels: L1 to L5 based on their capabilities and interaction logic.

We believe that the AI function of the current smartphone is stuck in the L2 stage and can only perform deterministic tasks, but we believe that the AI phone of the future will truly grow into an agent with a digital personality, which goes beyond the scope of understanding the task, has the ability to make independent decisions, and provides personalized services actively rather than passively.

How does AI mobile phone affect mobile phone manufacturers?

In a nutshell: mobile phone manufacturers enjoy volume and price dividends, hardware manufacturers benefit from product iteration and supply and demand improvement, and large model manufacturers accelerate their monetization.

Mobile phone manufacturers: The user base and traffic entrance create a moat, and the full-stack self-research can enjoy high profits

Looking back at the upgrade from 4G to 5G, we believe that AI mobile phones will benefit mobile phone manufacturers in terms of volume, price and profit.

Volume: AI mobile phones are expected to drive a new wave of replacement, and we estimate that AI mobile phone shipments will be about 380 million units in 2025, a year-on-year increase of 134%.

Price: The upgrade of AI mobile phones in terms of performance and user experience has surpassed the upgrade of 5G mobile phones relative to 4G mobile phones, and we judge that it is expected to replicate the previous round of price increases.

Pros: In the 5G era, most of the profits of pure mobile phone assembly manufacturers are eroded by the hardware side, the high cost of software and hardware of AI mobile phones, and mobile phone manufacturers with self-developed large models and self-developed chip capabilities occupy an integrated advantage and are expected to maintain profits.

Mobile phone manufacturers have the advantages of end consumer base and traffic entrance, and are more powerful in the process of cooperating with large model manufacturers. As the core carrier and entrance of the application layer, AI mobile phones will subvert the traditional mobile phone usage habits and lead a new round of industrial chain revolution. Under the wave of AI mobile phones, mobile phone manufacturers that grasp the first-mover advantage will benefit the most.